Manual masking in such a setup would be quite the task. If you’re using the latest version of Vegas Pro I recommend you give the “Smart Mask” plugin a try. This is a relatively new feature and I find it rather unreliable (haven’t yet tested on VP21), but you might have some better luck. I recommend checking out a video tutorial for it, it’s not super straight-forward.
If Smart Mask doesn’t work at all, then a slightly more manual process would be to add Bezeir Masking, create your mask, click on Motion Tracking and track through all of the footage. You might need multiple masks, like one for the head/face, one for shoulders/body, arms, etc.
Another approach is applying stabilization with “Video Stabilization” plugin. You might have to move it before “Pan/Crop” in the plugin chain. This approach might produce somewhat jittery/weirdly stabilized results though, needs a lot of fine-tuning.
Overall I’d say that tracking is still an area where Vegas Pro is somewhat behind. Many aspects of the process are surprisingly unintuitive (at least up to Vegas Pro 20), if possible I suggest looking into some AI tools/plugins to help with this task.